Where the Platform Debate Ended, and Where It’s Still Live 

Key takeaways

• The old build-or-buy debate has split into two questions with different answers. CFD and FX founders face a settled choice; exchange-traded founders face an open one.
• In CFD and FX, building a platform from scratch rarely repays the effort. Vendor platforms became the standard, and differentiation moved up the stack into spreads, research, pricing, and client experience.
• In options and futures, proprietary technology still wins. The depth a serious options trader expects, from risk graphs to multi-leg order entry, is hard to reach on a platform that looks identical to every competitor’s.
• A new exchange-traded operator has four routes: license as-is, commission a custom build on a proven base, build in-house, or assemble a hybrid from several vendors. The hybrid turns one decision into a layered set of vendor questions.
• Total cost of ownership decides who survives to year three. White-label brokers get caught by lock-in they did not price in; in-house builders get caught by maintenance, regulatory updates, market data fees, and engineering payroll that never stop.

Founders launching a brokerage in 2026 do not arrive at the platform decision the way founders did in 2006. The discussion that consumed broker startups twenty years ago has largely resolved itself in CFDs and FX, partially resolved itself in exchange-traded products, split into two very different conversations depending on which side of that line the new operator sits on, and acquired a much longer set of vendor-evaluation questions along the way. 

What used to be one question is now two questions with different answers. 

The CFD side: a debate that ended without a vote 

Through the early to mid-2000s, the CFD and FX retail space was overwhelmingly in-house. Established brokers built their own platforms because no one else’s tools were good enough yet. SaxoTrader, FXCM’s Trading Station, GFT’s DealBook 360, Oanda’s FXTrade, and CMC’s Marketmaker all served their own client bases, each one a major engineering investment. Devexperts was part of that wave; we built DealBook 360 for GFT Markets starting in 2002 and supported it for more than a decade, scaling it to 300+ white-labels across 120+ countries. 

By the mid-2010s, the picture had reorganized. Network effects on the trader side, where educators, signal authors, and algorithmic strategy communities all gathered around one platform family, made that family the default classroom for retail FX. Brokers who didn’t offer it were asking traders to learn a new tool before opening an account. Few traders did. The vendor option went from “an alternative” to “the standard,” and brokers shifted their differentiation upward, into spreads, research, education, partner programs, and customer experience. 

The result is that a new CFD brokerage today rarely has a meaningful debate over building from scratch. Time-to-market is measured in weeks against an industry where established players have a 15-year head start. Capital that would go into building a CFD platform tends to produce a product that is behind on launch, expensive to maintain, missing the community features traders expect, and saddled with mobile parity costs the founder did not budget for. For most new CFD operators, the choice is between white-label and a vendor-built custom platform that already owns the underlying technology. The fully custom in-house build sits as a theoretical third option that almost nobody picks anymore. 

This says nothing about the quality of the underlying technology. It says everything about the economics of platform competition in the CFD space, which has consolidated to a point where the cost of being different at the platform layer outweighs the value, and the cost of being similar at the platform layer is small enough to absorb. 

The exchange-traded side: the question is still live 

Equities, options, futures, and the prediction-market contracts that are starting to share the same operator base are a different story. 

Look at who runs proprietary technology in this segment today. Several of the most awarded retail options brokers in the US operate platforms that their own teams built from the ground up, and the awards travel with the technology rather than with white-label arrangements. The two most prominent active-trader brokerages by professional reputation, both with histories stretching back decades, run in-house systems that have anchored their differentiation since launch. One of the segment’s leading futures-focused platforms, with a user base in the millions, was built by the operator who runs it. An established US exchange group launched financial futures trading in early 2026 on in-house matching and execution technology. Proprietary technology is a defining feature of how the segment’s leaders got to where they are. 

Several things make the exchange-traded side different. 

The first is depth of differentiation. An options trader’s view of platform quality runs deep into the analytics: Greeks visualization, strategy roll mechanics, risk graphs, multi-leg order entry, live portfolio P&L, scanner tools, and exchange connectivity quality. Each of these is an engineering problem, and the brokers that win on options or futures tend to do so because their platform excels at one or two of them. That kind of differentiation is hard to achieve on a white-label foundation, where the most important screens look the same as those of every competitor. 

The second is the regulatory and connectivity weight. Exchange-traded markets in the US require FINRA, NFA, SEC, CFTC, or exchange-level memberships, depending on the segment, along with direct exchange connectivity, market data agreements with each exchange, and reporting infrastructure that handles the audit trail every regulator expects. The technical surface area is much larger than for an OTC CFD operation. Vendors who can deliver across all of that, at the depth a serious operator needs, are fewer in number. 

The third is institutional history. Many of the platform brands in this segment predate the modern vendor market by decades. They built when there was no alternative and kept building because the in-house investment became a strategic asset. New entrants face a different set of choices than legacy operators, but the existence of legacy operators sets the bar for what “good” looks like. 

What the new operator faces 

The decision tree for an exchange-traded launch is genuinely open. A new operator can do any of the following. 

They can license a platform from a vendor and run it as-is, with branding and configuration but no deep customization. Differentiation has to live somewhere other than the platform. 

They can commission a vendor to build something custom on top of a proven base, owning most of the resulting IP and avoiding the engineering recruitment problem. The model that produced DealBook 360 for the CFD space is the same model that produces custom platforms for exchange-traded brokers today; the Devexperts case study on a custom trading platform built to escape vendor lock-in is one example of how this path runs. Source-code ownership clauses, transfer agreements, and modular component licensing all become negotiable in ways they are not for white-label. 

They can build internally, staffing a financial software engineering team and accepting that the team has to keep growing forever. This is the path the established proprietary-platform brokers walked, but they did so from a position most new entrants cannot match: an incumbent client base, existing revenue, and the ability to absorb a five-year engineering ramp before launch. 

They can run hybrids. License the matching engine from one vendor, the charting library from another, build the trader-facing UX in-house, partner for market data, and integrate the back office. This is increasingly common, and it is where the question of vendor evaluation gets sharpest. The decision moves from “build or buy?” to a more layered set of questions about which pieces, from which vendors, on what contractual terms, with what integration plan, and how everything fits together over time. 

The cost trap that catches both sides 

The trap that catches founders on both sides is the same one Rado describes in his podcast appearance: total cost of ownership is the conversation no founder wants to have on day one, and it is also the conversation that decides who is still around in year three.

He frames the gap between the two sides this way:

The vendor’s concern is with stability, availability, and cost of the platform, whereas the broker is more concerned with the features the platform brings to improve the user experience.

On the CFD side, the trap usually shows up as a vendor lock-in surprise. The founder chooses white-label for speed, the platform proves harder to switch off than expected, and three years in, the broker finds that the differentiation that was supposed to happen above the platform never materialized. The fix is hard: re-platforming a live brokerage is one of the most expensive operations in the industry, and most brokers opt to expand their offerings. 

On the exchange-traded side, the trap shows up as the cost of “we’ll just build it.” Initial development is a fraction of the total. Maintenance, regulatory updates, market data fees, security patching, and engineering payroll all compound year over year to keep it all running. The people who built it are often the only ones who can maintain it, creating a hiring trap when they leave. Many operators who started by building end up either re-architecting around vendor components or selling to a larger acquirer who can absorb the engineering function. 

The second episode of the DXbrief podcast 

The other half of the conversation, what changes when you move from being inside a brokerage making this decision to consulting on it as a vendor, is the subject of our latest podcast with our colleague Radoslav Shalliev. He spent close to a decade as an account manager on the broker side before joining Devexperts, and his view of where the broker-vendor gap sits, plus why the cost and stability concerns that dominate vendor planning rarely surface during a broker’s launch, is worth listening to.

On what changed when he crossed from the broker side to the vendor side, he told us:

The biggest surprise for me was the amount of detail and focus the teams exhibit when they onboard a new customer and integrate the different pieces of technology.

You can watch the second episode of Devexperts’ podcast on our YouTube channel.

Contact us about your brokerage platform project to assess the right build-vs-buy model before deciding on cost, control, and integration.